AGE AND GROWTH OF THE CISCO 
303 
in either of the former populations. The difference in average length between fish 
taken in l}i- and 1 ^-inch-mesh nets was for example 5 millimeters in Trout Lake, 14 
millimeters in Muskellunge Lake, and 25 millimeters in Silver Lake. Further, the 
difference in average length between fish taken in l%- and 1%-inch-mesh nets in 
Muskellunge Lake was 10 millimeters while in the Silver Lake cisco this difference 
was 12 millimeters. 
Tables 74 and 75 show also that the average length for a particular age group 
varies considerably in samples taken by nets of different size mesh. Although cer- 
tain irregularities occur (as might be expected from the small size of the samples) 
there is a general tendency for the samples from larger mesh nets to show distinctly 
higher average lengths for a given age group than do the samples from nets of smaller 
mesh. Further examination shows, however, that this tendency does not in general 
affect the validity of the averages of age groups based on the combination of the 
samples of several nets. 
In the analysis of the data of tables 74 and 75 to establish the validity of the 
growth data for the Silver Lake cisco, it should be pointed out first that all ages above 
the II group in the 1931 sample taken in the new nets must be considered to have 
adequate representation throughout their entire length range. This is apparent from 
the fact that these older age groups are relatively scarce in the 1 ^-inch-net catch, 
while at the other extreme a 2%-inch-mesh net that was fished in the string failed to 
take any ciscoes at all. Since these average lengths of the older age groups from the 
new net sample can be considered reliable they can be used as a basis for the estima- 
tion of the reliability of the corresponding average lengths as determined from the 
old net samples. The fact that the average lengths for corresponding age groups of 
the 1930 and 1931 old net collections resemble each other on the whole more than 
either of them resembles the average lengths based on the 1931 new net collection 
suggests that the addition of the 1%- and 1%-inch nets may have increased the relia- 
bility of the Silver Lake samples. The differences between the corresponding aver- 
age lengths of the older age groups of the old and new net samples are not, however, 
of sufficient magnitude to invalidate the use of the earlier materials collected by the 
old gear. 
The II-group samples present a more difficult problem. The general scarcity 
of II-group individuals in the collections of the old gear would indicate that its length 
range may lie just below the range of efficiency of the lK-inch net, even though this 
size mesh was very efficient in Trout and Muskellunge Lakes in the taking of fish 
whose lengths were well below the observed average length (about 170 millimeters) 
for the Silver Lake II group. The 1931 sample with the new gear shows, however, 
that it is very unlikely that the Silver Lake II group has suffered serious selection by 
gear, as that selection affects average length, in any of the samples. The scarcity of 
the II-group fish in this sample taken in the new gear must be considered to show that 
this group in 1931 actually was less abundant than the neighboring age groups, for 
there is no reason to hold that the same nets that took 57 I-group and 30 Ill-group 
fish should fail to take II-group individuals in equal numbers provided they were 
equally abundant in the population. These facts together with the good agreement 
between the average lengths of the II-group samples as based on old and new gear 
samples make the retention of the II-group fish in the growth data advisable. 
Since there is no means of demonstrating the reliability of any of the I-group 
samples they must be considered, at the best, questionable. They were accordingly 
eliminated from the growth data. 
